<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Schneider, Thomas (NIH/NCI) [E] wrote on 2/6/2018 4:16 PM:<br>
<span style="white-space: pre-wrap; display: block; width: 98vw;">> Jeff:
>
>> I know that nondeterminism in network behaviors can cause
>> duplicated messages. However, I am seeing many more duplicate
>> emails in this mail list than any theory can explain. I want to
>> know if others are experiencing the same problem as I.
>
> I'm not seeing duplicate messages in texhax right now but have been
> puzzled about finding them sometimes in other contexts. What do you
> know about these?
</span>Our internet protocols and particularly our email transmission
protocols try to be error tolerant. This includes the potential to
retransmit a message from some store-and-forward station multiple
times. Since different paths can be followed by different
transmissions of the same message and the fact that the problem with
a transmission might not be the "sending" but loss of the "I've got
it" token, multiple copies can wend their way through the net's
structures by the same or different paths. These copies can arrive
at a common node near the same time or at wildly different times so
it is difficult to spot and remove duplicates in any reliable way
that doesn't involve enormous amounts of dedicated storage. Some
protocols such as TCP and FTP have duplicate spotting and removal
built in to them but most protocols such as those used for mail do
not (because there isn't enough information available at the
terminal to know when a duplicate isn't just identical and not a
duplicate). I ask my question on this mail list since I'm seeing
duplicates only here and not on the MikTex mail list for example.
I'm just looking for a quick clue.<br>
-- <br>
Jeff Barnett<br>
</body>
</html>